- Email: [email protected]
Prevalence of coronary heart disease in Turkish adults
International Journal of Cardiology, 93 (1993) 23-3 1 0 1993 Elsevier Scientific Publishers Ireland Ltd. All rights reserved. 0167-5273/93/$06.00
23
CARD10 01661
Prevalence of coronary heart disease in Turkish adults Altan Onat, Mustafa $. Senocak. Giinsel $urdum-Avci and Ender &nek Turkish Society of Cardiology. Istanbul, Turkey
(Received 13 October 1992; revision accepted 18 November 1992)
The prevalence of coronary heart disease was determined by a conducted survey in a random sample of 3689 subjects 20 years of age or older in 59 communities representing the Turkish adult population. Interview with a questionnaire, physical examination of the cardiovascular system and recording of a 12-lead ECG were performed. The latter was coded according to the Minnesota code. Expressed in age-adjusted rates (for 35-64 years), prevalence rates per 100 men were as follows: typical angina 3.7, atypical angina 0.9, electrocardiographic evidence of myocardial infarction and/or ischemia 3.7, any of the stated findings suggesting coronary heart disease 8. Women had a substantially higher rate of atypical angina, positive ECG findings and of any of the stated manifestations for coronary heart disease, whereas they had a significantly lower rate of Q/QS patterns as well as of a history of myocardial infarction. Based on a probability-related point score, age-adjusted clinical coronary heart disease was estimated to prevail in 5.8% of men and 5% of women (P > 0.4) in the sample of the Turkish population. The respective rates in urban residents was 6% and in rural resident 4.8%. Among participants diagnosed coronary heart disease, 63% presented the form of angina without infarction, 27% had evidence of myocardial infarction, 7% ‘silent myocardial ischemia’ and 3% cardiac failure alone. Key words: Angina pectoris; Coronary coronary disease
heart disease; Epidemiology;
Introduction Morbidity statistics in cardiovascular diseases are generally obtained from health institutions, health insurance returns and from populationbased surveys. Although standardized data regarding cardiovascular mortality in industrialized countries have been existing for many years, few countries have carried out morbidity surveys in the community [l]. Published data from the Seven Correspondence to: Professor Dr Altan Onat, Nisbetiye Cad. 37/24, Etiler 80630, Istanbul, Turkey.
Minnesota coding; Prevalence of
Countries [2], the Pooling Project [3] and the Paris Prospective Study [4], among others, contributed greatly to the knowledge on the incidence rate of the major subset of coronary heart disease in middle-aged men, namely on fatal and nonfatal myocardial infarction and death presumably due to coronary heart disease. Reliable prevalence data on cardiovascular morbidity in the developing countries are at best rare, and were inexistent in Turkey. This paper aims to describe the findings of a cross-sectional survey on the prevalence of coronary heart disease detected on the basis of history, physical examination and of related ECG tin-
24
dings obtained according to the Minnesota code in a representative sample of the Turkish adult population. An attempt was made to evaluate the prevalence of the most common form of coronary heart disease as well, namely angina pectoris without myocardial infarction. Differing from most previous surveys [5-71, the sample population in Turkey comprised women and men over a wide adult age range.
Subjects and Methods The survey on the Prevalence of Cardiac Disease and its Risk Factors in Adults in Turkey includes 3689 men and women 20 years of age and over residing in 59 different urban and rural communities scattered over all the seven geographical regions of Turkey. The criteria for selecting the urban and rural communities, participating subjects, the surveying teams and the steering committee, methods of data collection and the data obtained in the questionnaire have been presented in a separate report [8]. Briefly stated, a random sample of the Turkish adult population was surveyed with the purpose of determining the prevalence of heart diseases and the risk factors for coronary heart disease. The sample was representatively stratified for sex, age, geographic regions as well as the urban-rural distribution. For this purpose the number of participants in each gender, age group was predetermined in each urban and rural community in the geographic regions. The sizes of the selected communities conformed largely to the actual strata of township population in the country. Presumably, the distribution of ethnic groups in the sample reflected the actual one in Turkey, since the same proportion of subjects were sampled in all geographic regions. When each surveying team reached the sample community, they first obtained information about the socioeconomic distribution of the living quarters, then rang randomly preselected doors in the evening and gave appointments for an examination the next morning. The ratio of responders exceeded 85%. Since the survey sought to evaluate not only the prevalence of coronary heart disease, but rather the sex- and age-specific distribution of the risk factors among adults, the middle to upper
age group was not primarily targeted. The sample population had virtually an identical distribution of the Turkish adult male and female population. The age composition of the Turkish adult population was intentionally distorted in the sample solely in the age group 60-69 which should have comprised 3 10 screenees. However, 406 participants were surveyed in this group in order to reach a more meaningful statistical evaluation. An average of 62 persons per community were surveyed. Three teams were formed each consisting of two physicians at their fourth year of specialization in internal medicine and of a laboratory technician who was in postgraduate training in medical biology. The survey started on July 13th and ended late in September, 1990. Team members were given a training course for 2 days by the supervising staff to acquaint them with the technique of random sampling, the approach in the communities and other pertinent items. The task of each physician was to perform the interview, to examine the cardiovascular system as stipulated by the questionnaire, and to record an ECG. Each team had at its disposal apparatus including one scale with a measuring device for height, two anaeroid sphygmomanometers (Erka), one Reflotron instrument and one portable 3-channel electrocardiograph for 12 simultaneous leads (Cardiovit AT 3/l, Schiller). Logistics were provided by the Ministry of Health of the Turkish Republic. The presence of angina pectoris and exercise tolerance was specifically sought by directing preselected questions (comparable to the London School of Hygiene and Tropical Medicine Chest Pain Questionnaire [9]). The examiner’s impression was recorded regarding the absence of angina, presence of typical or atypical angina. Typical angina was defined as a chest pain located retrosternally or precordially which appeared while walking and disappeared within 2-15 min upon stopping or slowing down. Atypical angina was defined as chest pain located in parts of the chest other than those mentioned above, or when symptoms did not require stopping or slowing down, or when the pain did not disappear in 10 min. The questionnaire included the question as to whether a ‘heart attack’ had occurred (severe chest pain attack lasting for over 30 min) requiring treat-
25
ment with prolonged rest. A positive reply was suggestive of an old myocardial infarction which was considered definite either when this was observed in a subject with angina pectoris or in one exhibiting a Code 1 on the ECG. Physical examination included the palpation of the character, site, and size of cardiac impulses, ausculting the heart sounds and murmurs. The blood pressure was measured twice in the sitting position on the right arm at an interval of at least 3 min, and the mean value of two readings was used in the analysis. An average of 11 s of ECG recording was available in each subject. All ECGs were reviewed and coded according to the Minnesota code [lo] jointly by two experienced electrocardiographers. All collected survey data were checked by the supervising staff before being included in the data base with the purpose of assuring data quality and completeness. ECG evidence of myocardial infarction and ischemia is reported in this paper as Q-wave items (Code l. l-3), ST-segment or T-wave items (4.1-3, 5.1-3), or any of the previous items or Code 7.1 or 8.3 (left bundle-branch block or atria1 fibrillation or flutter). Isolated heart failure, in the absence of previously mentioned indicators of coronary heart disease, though an uncommon event, was recorded separately. The diagnosis was based on dyspnea at rest associated with an apex beat displaced to the left or on neck vein distension with hepatomegaly or peripheral edema. To test the inter-observer variability, 62 electrocardiograms from the survey comprising 17 tracings with ischemic ECG abnormalities coded by a referee were sent to the University of Kuopio, Finland, where our coding was compared with the coding results of two independent experts. Our prevalence for ‘ischemic’ findings was reported by the referee to be very close (closest of the three codings) to the ‘golden standard’. The prevalence of any of the previously stated manifestations or coronary heart disease was presented separately in men and women by age groups. In addition, a final category of clinical assessment of coronary heart disease was attempted based on a simple probability-related point score: coronary heart disease was considered definite and credited by one point in the presence any of the
following: (a) history of typical angina, (b) Minnesota Code 1.1-3 with a history of myocardial infarction, (c) history of myocardial revascularization and/or of a significant stenosis found at coronary arteriography, (d) manifestations of cardiac failure unaccompanied by a hypertensive, valvular or congenital aetiology. Coronary heart disease was considered suspect and credited by one-half point in the presence of either: (a) atypical angina pectoris, or (b) no angina in a subject 40 years of age or older having any of Codes 7.1, 4.1-2, 5.1-2, or Code 1.1-3 without a history of heart attack. As herein implied, minor ST-depression (Code 4.3) and minor T inversion (Code 5.3) not accompanied by symptoms or other manifestations by coronary heart disease was not comprised of this assessment to indicate disease in view of the low specificity in the general population. These minor ST-T changes when occurring in a hypertensive subject were judged to reflect hypertensive heart disease, the prevalence of which was evaluated separately though it will not be dealt with in this paper. The prevalence data are reported herein in loyear age-specific groups and age-adjusted. Age standardization was performed for age range 35-64 years with World Health Organization criteria, namely by utilizing population weights of 12131, 11131 and 8131 for the age groups 35-44, 45-54 and 55-64 years, respectively [ 111.
Results A. Prevalence of symptoms
The prevalences of typical angina pectoris, atypical angina and history of myocardial infarction by age groups in men and women are presented in Tables 1 and 2. Typical and atypical angina are mutually exclusive so that the sum of both prevalences indicate that of any type of angina pectoris. Age-adjusted typical and atypical angina in men were prevalent in 3.7% and 0.9%, respectively. The prevalence rates of both types rose with age but far more steeply (more than doubling with each age group) in typical than in atypical angina up to the age group 60-69. While being very infrequent (0.5%) in women, a
26 TABLE 1 Prevalence (in %) of symptoms and Minnesota codes suggestive of coronary heart disease in Turkish men by various age groups. Adjusted 20-29 years rate (n = 601) (35-64 years) No. Rate Typical angina Atypical angina History of MI Mn Code 1.1-3 Codes 4.1-3 & 5.1-3 Any of codes 1,4,5,7.1,8.3 Any item above Clinical CHD
30-39 years 424)
40-49 years (n = 296)
(n =
50-59 years 258)
(n =
60-69 years 206)
b 70 years (n = 82)
No.
Rate
No.
Rate
No.
Rate
No.
Rate
No.
Rate
0.7 0.9 0.5 0.7 0.5 0.7
7 2 2 4 3 5
2.4 0.7 0.7 1.4 1 1.7
13 1 8 4 7 12
5 0.4 3.1 1.6 2.7 4.7
25 6 14 13 22 28
12.1 2.9 6.8 6.3 10.7 13.6
1
1.2 1.2
2.4 1.5
14 9.5
4.7 3.2
28 20
10.9 7.8
45 35
21.8 17
(n =
3.7 0.9 1.8 1.8 2.5 3.7
0
2 2
0.2 0.3 0.3
3 4 2 3 2 3
8.0 5.8
3 0.5
0.5 0.1
10 6.5
I
0.2
0
I
1 0
1 5 5
1.2 6.1 6.1
8 3
9.8 3.7
MI, myocardial infarction; Mn, Minnesota; CHD, coronary heart disease.
history of myocardial infarction was noted in 1.8% (age-adjusted) in men. The related prevalence rates more than doubled with each age group to reach 6.8% in 60- to 69-year-old men. In women, age-adjusted typical and atypical angina prevailed in 3.5% and 2.0%, respectively. An abrupt rise from age 50 onwards was observed in the prevalence of typical angina in women, whereas that of atypical angina almost doubled with each age group. Heart failure not accompanied by chest pain symptoms or electrocar-
diographic signs of myocardial ischemia occurred only in three women past the age of 70. It was diagnosed also in a man each with typical and atypical angina and in three women (aged 59 or more) with typical angina. B. Prevalence of ECG findings Q/QS patterns suggesting an old myocardial infarction (Codes 1.1-3) were observed rarely in women (age-adjusted rate 0.25%), while they
TABLE 2 Prevalence (in %) of symptoms and Minnesota codes suggestive of coronary heart disease in Turkish women by various age groups. Adjusted 20-29 years rate (n = 586) (35-64 years) No. Rate Typical angina Atypical angina History of MI Mn Code 1.1-3 Codes 4.1-3 & 5.1-3 Any of codes 1,4,5,7.1,8.3 Any item above Clinical CHD
3.5 2.0 0.5 0.25 3.0 4.9
0 2 0 0 0 0
9.6 5.0
2
I
0.3
0.3 0.2
30-39 years (n = 411)
40-49 years (n = 291)
50-59 years (n = 252)
60-69 years (n = 199)
2 70 years (n = 83)
No.
Rate
No.
Rate
No.
Rate
No.
Rate
No.
Rate
4 2 0 0 6 8
1 0.5
0.7 1.4
17 5 3
7.0 4.0 1.5 2.0 8.0 11.1
4.8 1.2
12 15
14 8 3 4 16 22
1
1.4 2.7
6.7 2 1.2 0.4 4.8 6
4
1.5 1.9
2 4 0 0 4 8
0 0 11 15
13.3 18.1
11 6
2.7 1.5
13 4
4.5 1.4
34 21
13.5 8.3
41 22
20.6 11.1
21 10
25.3 12.0
1
MI, myocardial infarction; Mn, Minnesota; CHD, coronary heart disease.
21
prevailed in 1.8% of men. A steep age gradient existed more than doubling with each age group up to age 69 in men. The trend was valid in women as well in whom Code 1 was not met under the age of 50 years. Of 31 subjects (26 men and five women) coded with Q/QS patterns, 15 revealed definite evidence of coronary heart disease, two were considered to have cardiomyopathy, and 13 were asymptomatic so they were classified under suspected coronary heart disease (one-half point). Virtually half of the Q/QS patterns were coded 1.1, the remainder fell predominantly to medium Qwave items. Code 4 and/or 5 items suggesting myocardial ischemia prevailed in 2.5% of men and 3% of women (age-adjusted). The rate of these findings roughly doubled with each age group in men and rose by over 70% in women with each decade. ST depression findings were observed in a total of 29 persons, the majority having minor ST depression (Code 4.3). Except for three subjects with Code 4.3 alone, all were regarded as having definite or suspected organic heart disease. T-wave inversion or flattening was fairly common, having been recorded in 90 persons, especially among those aged 50 and over. The great majority of them had intermediate or minor T-wave changes (Code 5.2 and 5.3). Fifteen subjects were considered not to have organic heart disease, 22 had suspected coronary (or other) cardiac disease. Of 53 subjects with definite cardiac disease, 3 1 were classified as coronary, 16 as hypertensive disease, and the remainder as rheumatic, car pulmonale or cardiomyopathy. Left bundle-branch block encountered in 13 subjects (five men and eight women) and atria1 fibrillation (or flutter) in 14 persons (six men and eight women) were considered to indicate the presence of organic disease. Persons showing any of the previously mentioned Q or ST-T changes or exhibiting left bundle-branch block or atria1 fibrillation were classified separately (Tables 1 and 2). Age-adjusted prevalence rate of this category was 3.7% in men and 4.9% in women. C. Prevalence of coronary heart disease The age-adjusted prevalence rate of any clinical
Is3 Prevalence % Male
1 Prevalence % Female
% 1 El 16 14 12 10 8 6 4 2 0 20-29
30-39
40.49
50-59
60-69
270
Age Groups
Fig. 1. Prevalence
of coronary heart disease in Turkish by age groups.
adults
or electrocardiographic manifestation of heart disease described in Tables I and 2 were 8% and 9.6% in men and women, respectively. Since this definition probably includes a substantial number of normal subjects as well as patients with noncoronary disease such as hypertensive heart disease or cardiomyopathy, a clinically assessed prevalence of coronary heart disease was sought to be specified. These age-adjusted rates were 5.8% in Turkish men and 5% in women (P > 0.4) (Tables 1 and 2). A steep age gradient existed in the coronary heart disease prevalence in both genders, multiplying by 2.2 to five times in each life decade in women, and by roughly three times in men before declining from age 70 onwards. TABLE
3
Prevalence (in o/o) of coronary in various age groups. Age groups
All adults 20-29 years 30-39 years 40-49 years 50-59 years 60-69 years z 70 years
heart disease in Turkish
Men
Women
A
B
Rate
A
B
Rate
59 0 3 8 18 29 1
31 1 7 3 4 12 4
4 0.1 1.5 3.2 7.8 17.2 3.7
45 0 4 2 18 14 7
38 2 4 4 6 17 5
3.5 0.2 1.5 1.4 8.3 11.3 11.5
Column A, Number Column B, Number point given).
adults
of subjects with definite CHD. of subjects with suspect CHD
(one-half
28
D. Prevalence of coronary heart disease by (urban vs. rural) residence The number of persons with a definite or suspected diagnosis of coronary heart disease and their prevalences by various age groups are shown in Table 3. Age-adjusted prevalence rate of coronary heart disease in Turkish men was 5.8% in the general population, 5.7% in the urban and 6.1% in the rural population (P > 0.4). The respective rates in Turkish women were 5,6.4 and 3.5% (P c 0.06 between urban and rural sections).
E. Distribution of the presenting form of coronary heart disease Of 139 participants with clinical coronary heart disease, 87 (63%) presented angina pectoris without myocardial infarction. The latter comprised 70 subjects with typical and 34 with atypical angina (credited by one-half). In 32 patients with or without angina there was definite evidence of myocardial infarction, whereby nine persons without angina presented a history of myocardial infarction and/or Code 1. In further 10 subjects without angina Code 1 existed. Thus 27% of participants with coronary heart disease showed evidence of myocardial infarction. Silent myocardial ischemia as judged from Code 4 and/or Code 5 items in symptom-free subjects formed 7% of those having coronary heart disease, while heart failure alone was considered the presenting form of the disease in 3%.
F. Summary of distribution of some risk factors Following data are provided in order to give an idea as to the significance of underlying risk factors. Mean plasma concentrations adjusted for the age range 35-64 years were 4.8 mmoVl(185 mg/dl) in men and 5 mmol/l (192 mg/dl) in women [8]. Hypercholesterolemia (L 6/2 mmol/l (240 mg/dl) prevailed among 10% of men and 11.7% of women. Roughly two-thirds of these men and women had levels of plasma total cholesterol in excess of 6.5 mmol/l. While age-adjusted mean plasma triglyceride concentration was 1.69 mmol/l (150.2 mg/dl) in
men and 1.46 mmol/l (129.2 mg/dl) in women, hypertriglyceridemia ( 2 2.26 mmol/l, 200 mg/dl) was observed in 19.3% of men and 14.7% of women in the decades spanning the ages 40-69 years [8]. Hypertensive subjects were defined to have a systolic (~160 mmHg), diastolic (~95 mmHg) pressure, or both, or who reported to take antihypertensive medication. With an overall ageadjusted odds ratio of 2.86, hypertension constituted the most important risk factor in the sample population [ 121. It prevailed among 17% of men and 23.4% of women, when standardized for ages 35-64 years. Again when age-adjusted, 49% of Turkish men smoked daily more than 10 cigarettes, and 8% smoked up to 10 cigarettes. The respective figures for smoking women were 9% and 5.5%. Turkish women, but not men, distinctly tended to be obese: 36.5% of age-adjusted women had a body mass index ~30.0 kg/m2. Exhibiting an odds ratio of 1.76, obesity was an additional significant risk factor in women.
Discussion Coronary heart disease prevalence rates of populations obtained in epidemiologic studies have some shortcomings to be compared with each other due to a variety of reasons. Nonetheless in this study prevalences of typical and atypical angina pectoris, partial indicators of coronary heart disease, can fairly readily be compared in view of commonly adopted criteria for these symptoms. Typical angina was marginally more common in men than in women (age-adjusted 3.7% vs. 3.5%), while atypical angina was elicited twice as commonly in women (2.0% vs. 0.9% in men). In agreement with available knowledge, the relatively high prevalence of angina in women has been attributed to a lower pain tolerance [ 131, or to a greater socio-psychological acceptability in Western cultures [ 141, or to a higher proportion of women with less severe coronary heart disease [15]. The prevalence rate in Turkish women was two-thirds of that of Finnish women with respect to typical angina and one-quarter as regards atypical angina [ 151. Our rates in men were close
29
to those of Finnish men in respect with typical angina, while they constituted one-tenth as regards atypical angina [15]. Criteria for atypical angina may have been applied in a narrower sense in the present study. For further comparison one may add that typical angina prevailed in 5.5% of men aged 50-59 years in the Whitehall Study [16], in 5.8% of men aged 40-59 in the British Regional Heart Study [6] and in 5.7% of those aged 47-53 in the second Goteborg study [ 171. The prevalence of severe prolonged chest pain attack has varied widely between populations, such as for men aged 50-59 years from that of 2.4-9% [ 18,151 and data on this point might prove still less reliable as an indication of an old myocardial infarction, consequent to a lower level of health awareness among Turkish adults. Therefore, a stricter definition of a history of myocardial infarction was used here which included prolonged rest following a chest pain lasting onehalf hour or more. This provided an infarct prevalence in men that represented one-half that of typical angina. This ratio in men might somewhat underestimate the proportion of myocardial infarction in the spectrum of manifestations of coronary heart disease. Among Q/QS patterns in the ECGs, the overwhelming majority of subjects with Code 1 comprised Code 1.1 and 1.2 items. The age-adjusted prevalence of Code 1 in Turkish men was 1.8%, a figure though approaching the corresponding rate of 2.2% obtained by dstiir et al. [ 191, is half as low as that in the Framingham [20], Whitehall [ 161and Finnish [ 151 studies. The corresponding prevalence in women aged 50-59 years was approximately half that of men in three of the just mentioned studies, while Q/QS patterns prevailed in Turkish women at one-fifth the rate of male participants. The low rate of both the history of myocardial infarction and of Code 1 in Turkish women suggests that myocardial infarction predilects women much less and men more strongly than in several studies from industrialized nations. For example, among randomly selected residents of Rochester, Minnesota, among cases of coronary heart disease those with myocardial infarction constituted 54% in men and 33% in women [21]. ST depressions prevailed among the Finnish
sample population in about 3% and any T-wave changes (Code 5.1-3) in about 7.5% when similar age-adjustment is performed as in the present survey. These rates are comparable to those in several other surveys [6,8,19], while the rate for T-wave inversions among Turkish men (2.5%) was about one-third. The policy in this study of crediting half a point to asymptomatic subjects with left bundle-branch block seems supported by the experience in the Framigham study [22] in which only 11% among subjects with a newly acquired left bundle-branch block remained free of clinically apparent cardiovascular abnormalities throughout a period of 12 years’ observation: namely, since merely three of our 13 subjects with left bundle-branch block were symptom-free, 12% (1.5:13) of all those with left bundle-branch block was excluded from a diagnosis of disease. When any clinical and electrocardiographic manifestation suggestive of coronary heart disease is considered, 8% of men and 9.6% of Turkish women (age-adjusted) can be classified as such. This can be compared with the findings in the British Regional Heart Study [6] in which 24.7% of men aged 40-59 years had some evidence of coronary heart disease by survey questionnaire or ECG or both. Criteria as broad as these probably inflate the true prevalence of coronary heart disease. This is true in men, but more so in women in whom, in addition to atypical angina, minor ST-T changes are often false positively recorded [23]. Lability in the autonomic nervous system and constitutional differences in women such as a more backward directed T vector in the horizontal plane have been implicated to underly this phenomenon [24]. Therefore, a point-score system as devised herein which credits clinical and ECG findings with a low reliability by one-half point would seem to lead closer to the real prevalence of coronary heart disease. Our criteria for coronary heart disease essentially excluded isolated Codes 4.3 and 5.3 (not associated with angina) and counted half the number of persons exhibiting atypical angina or isolated electrocardiographic ‘myocardial ischemia’ of greater than minor severity. The latter approach took into account the existence of ECG changes of a noncoronary origin. Evidently, the
30
value of this assessment based on clinical judgment needs to be demonstrated in a prospective study. Nevertheless, it is provided here not as a substitute but as a supplement to the evaluation of coronary heart disease indicators obtained by standard epidemiological methods. Based on the point score, Turkish subjects, adjusted for age range 35-64 years, have a coronary heart disease prevalence of 5.8% in men and 5% in women. Overall, the age-adjusted prevalence among the urban population of 6% exceeded that of the rural population (4.8%), and this difference stemmed from women. Higher coronary heart disease prevalence data in urban residents in our study was associated with a mean total cholesterol level in the urban population (age-adjusted 193 mgdl in men vs. 182.7 mg/dl in women) which was significantly higher by 10.3 mg/dl (0.27 mmol/l) than in the rural population [8]. Though rates of symptoms and of ischemic ECG findings rose substantially with increasing age in women across all age groups and in men until age 69, lower prevalence rates of coronary heart disease were observed in men aged 70 or over implying a healthier status. Since the number of subjects comprised in this age category was comparatively low, its confidence interval is likely to be large. The low prevalence rate in the oldest men might either represent a sampling bias in this category, or it might reflect a natural selection, namely the elimination of most Turkish men with coronary disease by death before reaching the age of 70. The latter possibility is in keeping with the fact that mean age at death for Turkish men is 63.9 in contrast to 69 in women. A prevalence rate for coronary heart disease by the clinically assessed criteria in women as close to men (by merely 13% lower), or even higher than men by combined use of angina and ECG findings with standard survey methods is not an observation commonly made and needs commenting. In addition to the previously discussed factors tending to augment the coronary heart disease prevalence in women in general, a significantly higher prevalence of obesity, hypertension and of diabetes in Turkish women than in men may account for the finding of a relatively high prevalence rate of coronary heart disease in women.
In an epidemiologic cross-sectional study aimed at eliciting the prevalence of coronary heart disease angina, severe chest pain attack and ECG findings are the major tools in detection. In this survey a history of angina and/or myocardial infarction had 67% of weighting in the diagnosis, a history of bypass surgery or of angiographic tindings had a 3% weight and a similar share was by manifestations of heart failure at physical examination. The ECG provided a diagnosis in 27% of instances; of this proportion sequelae of myocardial infarction and symptomatic myocardial ischemia comprised 20%, while ‘silent’ ischemia had 7% weight. (In instances of fullillment of combined (history and ECG) half shares were allotted to both methods in the detection.) Thus two to threefold emphasis was given in this study in the diagnosis of coronary heart disease on isolated chest pain symptoms than on the presence of isolated ECG findings, an approach the correctness of which was advocated by Reunanen and associates
v51. In conclusion, this study provided a comparatively low prevalence of indicators of coronary heart disease in Turkish adults which was associated with generally low concentrations of serum total cholesterol. The prevalence rate was slightly but not significantly higher in the urban than in the rural sample population, and only marginally lower in Turkish women than in men.
Acknowledgements We appreciate the dedicated works of Dr Y. Goziikara, Dr M. Ivler, Dr Y. Karaaslan, Dr J. iSzisik, Dr F. Tabak, the coworkers in the survey teams, and that of V. Taskin and 6. oz in the data processing. The support of Professor R. &can, president of the Turkish Society of Cardiology is gratefully acknowledged. The fruitful suggestions by Professor K. Pyorala (Kuopio, Finland) in the preparation of the manuscript is deeply appreciated.
References 1
Uemura K. International trends in cardiovascular disease in the elderly. Eur Heart J 1988; 9 (Suppl D): l-8.
31 2 3
4
5
6
7
8
9
10
11
Keys A, editor. Coronary heart disease in seven countries. Circulation 1970; 41 (Suppl 1): l-195. The Pooling Project Research Group. Relationship of blood pressure, serum cholesterol, smoking habit, relative weight and ECG abnormalities to incidence of major coronary event: final report of the Pooling Project. J Chron Dis 1978; 31: 201-306. Ducimetibre P, Richard JL, Cambien F, Rakotovao R, Claude JR. Coronary heart disease in middle-aged Frenchmen: Comparisons between Paris Prospective Study, and Pooling Project. Lancet 1980; 1: 1346-50. Reid DD, Hamilton PJS, McCartney P, Rose G, Jarret RJ, Keen H. Smoking and other risk factors for coronary heart disease in British civil servants. Lancet 1976; ii: 979-83. Shaper AG, Cook DG, Walker M, MacFarlane PW. Prevalence of ischaemic heart disease in middle-aged British men. Br Heart J 1984; 51: 595-605. Bainton D, Baker IA, Sweetnam PM, Yarnell JWG, Elwood PC. Prevalence of ischaemic heart disease: the Caerphilly and Speedwell surveys. Br Heart J 1988; 59: 201-206. Onat A, Surdum-Avci G, Senocak M, Grnek E, Goziikara Y. Serum lipids and their interrelation in Turkish adults. J Epidemiol Commun Health 1992; 46: 470-476. Rose G, McCartney P, Reid DD. Selfadministration of a questionnaire on chest pain and intermittent claudication. Br J Prev Sot Med 1977; 31: 42-43. Rose GA, Blackburn H, Gillum RF, Prineas RJ. Cardiovascular Survey Methods, 2nd edn. Geneva, WHO 1982; 124-127. The WHO MONICA Project. Geographical variation in
the major risk factors of coronary heart disease in men and women aged 35-64 years. World Health Stat Q 1988; 41: 115-140. 12 Onat A, Senocak MS. Prevalence and clustering of risk factors for coronary heart disease in Turkish adults, and the associated relative risks. Arch Turk Sot Cardiol 1992; 20: 129-136.
16
17
18
19
20
21
22
23
24
Woodrow KM, Friedman GD, Siegelaub AB, Cohen MF. Pain tolerance: differences according to age, sex and race. Psychosom Med 1972; 34: 548-556. Nathanson CA. Illness and the feminine role: a theoretical review. Sot Sci Med 1975; 9: 57-62. Reunanen A, Aromaa A, Pyiirala K. Punsar S, Maatela J, Knekt P. The Social Insurance Institution’s Coronary Heart Disease Study. Helsinki: Social Insurance Institution, Finland. 1983; 120. Rose G, Baxter PJ, Reid DD, McCartney P. Prevalence and prognosis of electrocardiographic findings in middleaged men. Br Heart J 1978; 40: 636-643. Hagrnan M, Jonsson D, Wilhelmsen L. Prevalence of angina pectoris and myocardial infarction in a general population sample of Swedish men. Acta Med Stand 1977; 201: 571-577. Gyntelberg F. Physical fitness and coronary heart disease. Male residents in Copenhagen aged 40-59. Dan Med Bull 1973; 20: l-4. &tBr E, Schnor P. Jensen G. Nyboe J, Tybjaerg Hansen A. Electrocardiographic findings and their association with mortality in the Copenhagen City Heart Study. Eur Heart J 1981; 2: 317-328. Higgins ITT, Kannel WB, Dawber TR. The electrocardiogram in epidemiological studies: reproducibility, validity and international comparison. Br J Prev Sot Med 1965; 53-68. Philips SJ. Whisnant JP, O’Fallon WM, Frye RL. Prevalence of cardiovascular disease and diabetes mellitus in residents of Rochester, Minnesota. Mayo Clin Proc 1990; 65: 344-359. Schneider JF, Thomas HE Jr, Kreger BE et al. Newly acquired left bundle branch block: the Framingham study. Ann Intern Med 1979; 90: 303-310. Joy M, Trump DW. Significance of minor ST segment and T wave changes in the resting electrocardiogram of asymptomatic subjects. Br Heart J 1981; 45: 48-55. Simonson E. Differentiation between normal and abnormal in electrocardiography. St Louis. MO: C.V. Mosby Co, 1961.
23
CARD10 01661
Prevalence of coronary heart disease in Turkish adults Altan Onat, Mustafa $. Senocak. Giinsel $urdum-Avci and Ender &nek Turkish Society of Cardiology. Istanbul, Turkey
(Received 13 October 1992; revision accepted 18 November 1992)
The prevalence of coronary heart disease was determined by a conducted survey in a random sample of 3689 subjects 20 years of age or older in 59 communities representing the Turkish adult population. Interview with a questionnaire, physical examination of the cardiovascular system and recording of a 12-lead ECG were performed. The latter was coded according to the Minnesota code. Expressed in age-adjusted rates (for 35-64 years), prevalence rates per 100 men were as follows: typical angina 3.7, atypical angina 0.9, electrocardiographic evidence of myocardial infarction and/or ischemia 3.7, any of the stated findings suggesting coronary heart disease 8. Women had a substantially higher rate of atypical angina, positive ECG findings and of any of the stated manifestations for coronary heart disease, whereas they had a significantly lower rate of Q/QS patterns as well as of a history of myocardial infarction. Based on a probability-related point score, age-adjusted clinical coronary heart disease was estimated to prevail in 5.8% of men and 5% of women (P > 0.4) in the sample of the Turkish population. The respective rates in urban residents was 6% and in rural resident 4.8%. Among participants diagnosed coronary heart disease, 63% presented the form of angina without infarction, 27% had evidence of myocardial infarction, 7% ‘silent myocardial ischemia’ and 3% cardiac failure alone. Key words: Angina pectoris; Coronary coronary disease
heart disease; Epidemiology;
Introduction Morbidity statistics in cardiovascular diseases are generally obtained from health institutions, health insurance returns and from populationbased surveys. Although standardized data regarding cardiovascular mortality in industrialized countries have been existing for many years, few countries have carried out morbidity surveys in the community [l]. Published data from the Seven Correspondence to: Professor Dr Altan Onat, Nisbetiye Cad. 37/24, Etiler 80630, Istanbul, Turkey.
Minnesota coding; Prevalence of
Countries [2], the Pooling Project [3] and the Paris Prospective Study [4], among others, contributed greatly to the knowledge on the incidence rate of the major subset of coronary heart disease in middle-aged men, namely on fatal and nonfatal myocardial infarction and death presumably due to coronary heart disease. Reliable prevalence data on cardiovascular morbidity in the developing countries are at best rare, and were inexistent in Turkey. This paper aims to describe the findings of a cross-sectional survey on the prevalence of coronary heart disease detected on the basis of history, physical examination and of related ECG tin-
24
dings obtained according to the Minnesota code in a representative sample of the Turkish adult population. An attempt was made to evaluate the prevalence of the most common form of coronary heart disease as well, namely angina pectoris without myocardial infarction. Differing from most previous surveys [5-71, the sample population in Turkey comprised women and men over a wide adult age range.
Subjects and Methods The survey on the Prevalence of Cardiac Disease and its Risk Factors in Adults in Turkey includes 3689 men and women 20 years of age and over residing in 59 different urban and rural communities scattered over all the seven geographical regions of Turkey. The criteria for selecting the urban and rural communities, participating subjects, the surveying teams and the steering committee, methods of data collection and the data obtained in the questionnaire have been presented in a separate report [8]. Briefly stated, a random sample of the Turkish adult population was surveyed with the purpose of determining the prevalence of heart diseases and the risk factors for coronary heart disease. The sample was representatively stratified for sex, age, geographic regions as well as the urban-rural distribution. For this purpose the number of participants in each gender, age group was predetermined in each urban and rural community in the geographic regions. The sizes of the selected communities conformed largely to the actual strata of township population in the country. Presumably, the distribution of ethnic groups in the sample reflected the actual one in Turkey, since the same proportion of subjects were sampled in all geographic regions. When each surveying team reached the sample community, they first obtained information about the socioeconomic distribution of the living quarters, then rang randomly preselected doors in the evening and gave appointments for an examination the next morning. The ratio of responders exceeded 85%. Since the survey sought to evaluate not only the prevalence of coronary heart disease, but rather the sex- and age-specific distribution of the risk factors among adults, the middle to upper
age group was not primarily targeted. The sample population had virtually an identical distribution of the Turkish adult male and female population. The age composition of the Turkish adult population was intentionally distorted in the sample solely in the age group 60-69 which should have comprised 3 10 screenees. However, 406 participants were surveyed in this group in order to reach a more meaningful statistical evaluation. An average of 62 persons per community were surveyed. Three teams were formed each consisting of two physicians at their fourth year of specialization in internal medicine and of a laboratory technician who was in postgraduate training in medical biology. The survey started on July 13th and ended late in September, 1990. Team members were given a training course for 2 days by the supervising staff to acquaint them with the technique of random sampling, the approach in the communities and other pertinent items. The task of each physician was to perform the interview, to examine the cardiovascular system as stipulated by the questionnaire, and to record an ECG. Each team had at its disposal apparatus including one scale with a measuring device for height, two anaeroid sphygmomanometers (Erka), one Reflotron instrument and one portable 3-channel electrocardiograph for 12 simultaneous leads (Cardiovit AT 3/l, Schiller). Logistics were provided by the Ministry of Health of the Turkish Republic. The presence of angina pectoris and exercise tolerance was specifically sought by directing preselected questions (comparable to the London School of Hygiene and Tropical Medicine Chest Pain Questionnaire [9]). The examiner’s impression was recorded regarding the absence of angina, presence of typical or atypical angina. Typical angina was defined as a chest pain located retrosternally or precordially which appeared while walking and disappeared within 2-15 min upon stopping or slowing down. Atypical angina was defined as chest pain located in parts of the chest other than those mentioned above, or when symptoms did not require stopping or slowing down, or when the pain did not disappear in 10 min. The questionnaire included the question as to whether a ‘heart attack’ had occurred (severe chest pain attack lasting for over 30 min) requiring treat-
25
ment with prolonged rest. A positive reply was suggestive of an old myocardial infarction which was considered definite either when this was observed in a subject with angina pectoris or in one exhibiting a Code 1 on the ECG. Physical examination included the palpation of the character, site, and size of cardiac impulses, ausculting the heart sounds and murmurs. The blood pressure was measured twice in the sitting position on the right arm at an interval of at least 3 min, and the mean value of two readings was used in the analysis. An average of 11 s of ECG recording was available in each subject. All ECGs were reviewed and coded according to the Minnesota code [lo] jointly by two experienced electrocardiographers. All collected survey data were checked by the supervising staff before being included in the data base with the purpose of assuring data quality and completeness. ECG evidence of myocardial infarction and ischemia is reported in this paper as Q-wave items (Code l. l-3), ST-segment or T-wave items (4.1-3, 5.1-3), or any of the previous items or Code 7.1 or 8.3 (left bundle-branch block or atria1 fibrillation or flutter). Isolated heart failure, in the absence of previously mentioned indicators of coronary heart disease, though an uncommon event, was recorded separately. The diagnosis was based on dyspnea at rest associated with an apex beat displaced to the left or on neck vein distension with hepatomegaly or peripheral edema. To test the inter-observer variability, 62 electrocardiograms from the survey comprising 17 tracings with ischemic ECG abnormalities coded by a referee were sent to the University of Kuopio, Finland, where our coding was compared with the coding results of two independent experts. Our prevalence for ‘ischemic’ findings was reported by the referee to be very close (closest of the three codings) to the ‘golden standard’. The prevalence of any of the previously stated manifestations or coronary heart disease was presented separately in men and women by age groups. In addition, a final category of clinical assessment of coronary heart disease was attempted based on a simple probability-related point score: coronary heart disease was considered definite and credited by one point in the presence any of the
following: (a) history of typical angina, (b) Minnesota Code 1.1-3 with a history of myocardial infarction, (c) history of myocardial revascularization and/or of a significant stenosis found at coronary arteriography, (d) manifestations of cardiac failure unaccompanied by a hypertensive, valvular or congenital aetiology. Coronary heart disease was considered suspect and credited by one-half point in the presence of either: (a) atypical angina pectoris, or (b) no angina in a subject 40 years of age or older having any of Codes 7.1, 4.1-2, 5.1-2, or Code 1.1-3 without a history of heart attack. As herein implied, minor ST-depression (Code 4.3) and minor T inversion (Code 5.3) not accompanied by symptoms or other manifestations by coronary heart disease was not comprised of this assessment to indicate disease in view of the low specificity in the general population. These minor ST-T changes when occurring in a hypertensive subject were judged to reflect hypertensive heart disease, the prevalence of which was evaluated separately though it will not be dealt with in this paper. The prevalence data are reported herein in loyear age-specific groups and age-adjusted. Age standardization was performed for age range 35-64 years with World Health Organization criteria, namely by utilizing population weights of 12131, 11131 and 8131 for the age groups 35-44, 45-54 and 55-64 years, respectively [ 111.
Results A. Prevalence of symptoms
The prevalences of typical angina pectoris, atypical angina and history of myocardial infarction by age groups in men and women are presented in Tables 1 and 2. Typical and atypical angina are mutually exclusive so that the sum of both prevalences indicate that of any type of angina pectoris. Age-adjusted typical and atypical angina in men were prevalent in 3.7% and 0.9%, respectively. The prevalence rates of both types rose with age but far more steeply (more than doubling with each age group) in typical than in atypical angina up to the age group 60-69. While being very infrequent (0.5%) in women, a
26 TABLE 1 Prevalence (in %) of symptoms and Minnesota codes suggestive of coronary heart disease in Turkish men by various age groups. Adjusted 20-29 years rate (n = 601) (35-64 years) No. Rate Typical angina Atypical angina History of MI Mn Code 1.1-3 Codes 4.1-3 & 5.1-3 Any of codes 1,4,5,7.1,8.3 Any item above Clinical CHD
30-39 years 424)
40-49 years (n = 296)
(n =
50-59 years 258)
(n =
60-69 years 206)
b 70 years (n = 82)
No.
Rate
No.
Rate
No.
Rate
No.
Rate
No.
Rate
0.7 0.9 0.5 0.7 0.5 0.7
7 2 2 4 3 5
2.4 0.7 0.7 1.4 1 1.7
13 1 8 4 7 12
5 0.4 3.1 1.6 2.7 4.7
25 6 14 13 22 28
12.1 2.9 6.8 6.3 10.7 13.6
1
1.2 1.2
2.4 1.5
14 9.5
4.7 3.2
28 20
10.9 7.8
45 35
21.8 17
(n =
3.7 0.9 1.8 1.8 2.5 3.7
0
2 2
0.2 0.3 0.3
3 4 2 3 2 3
8.0 5.8
3 0.5
0.5 0.1
10 6.5
I
0.2
0
I
1 0
1 5 5
1.2 6.1 6.1
8 3
9.8 3.7
MI, myocardial infarction; Mn, Minnesota; CHD, coronary heart disease.
history of myocardial infarction was noted in 1.8% (age-adjusted) in men. The related prevalence rates more than doubled with each age group to reach 6.8% in 60- to 69-year-old men. In women, age-adjusted typical and atypical angina prevailed in 3.5% and 2.0%, respectively. An abrupt rise from age 50 onwards was observed in the prevalence of typical angina in women, whereas that of atypical angina almost doubled with each age group. Heart failure not accompanied by chest pain symptoms or electrocar-
diographic signs of myocardial ischemia occurred only in three women past the age of 70. It was diagnosed also in a man each with typical and atypical angina and in three women (aged 59 or more) with typical angina. B. Prevalence of ECG findings Q/QS patterns suggesting an old myocardial infarction (Codes 1.1-3) were observed rarely in women (age-adjusted rate 0.25%), while they
TABLE 2 Prevalence (in %) of symptoms and Minnesota codes suggestive of coronary heart disease in Turkish women by various age groups. Adjusted 20-29 years rate (n = 586) (35-64 years) No. Rate Typical angina Atypical angina History of MI Mn Code 1.1-3 Codes 4.1-3 & 5.1-3 Any of codes 1,4,5,7.1,8.3 Any item above Clinical CHD
3.5 2.0 0.5 0.25 3.0 4.9
0 2 0 0 0 0
9.6 5.0
2
I
0.3
0.3 0.2
30-39 years (n = 411)
40-49 years (n = 291)
50-59 years (n = 252)
60-69 years (n = 199)
2 70 years (n = 83)
No.
Rate
No.
Rate
No.
Rate
No.
Rate
No.
Rate
4 2 0 0 6 8
1 0.5
0.7 1.4
17 5 3
7.0 4.0 1.5 2.0 8.0 11.1
4.8 1.2
12 15
14 8 3 4 16 22
1
1.4 2.7
6.7 2 1.2 0.4 4.8 6
4
1.5 1.9
2 4 0 0 4 8
0 0 11 15
13.3 18.1
11 6
2.7 1.5
13 4
4.5 1.4
34 21
13.5 8.3
41 22
20.6 11.1
21 10
25.3 12.0
1
MI, myocardial infarction; Mn, Minnesota; CHD, coronary heart disease.
21
prevailed in 1.8% of men. A steep age gradient existed more than doubling with each age group up to age 69 in men. The trend was valid in women as well in whom Code 1 was not met under the age of 50 years. Of 31 subjects (26 men and five women) coded with Q/QS patterns, 15 revealed definite evidence of coronary heart disease, two were considered to have cardiomyopathy, and 13 were asymptomatic so they were classified under suspected coronary heart disease (one-half point). Virtually half of the Q/QS patterns were coded 1.1, the remainder fell predominantly to medium Qwave items. Code 4 and/or 5 items suggesting myocardial ischemia prevailed in 2.5% of men and 3% of women (age-adjusted). The rate of these findings roughly doubled with each age group in men and rose by over 70% in women with each decade. ST depression findings were observed in a total of 29 persons, the majority having minor ST depression (Code 4.3). Except for three subjects with Code 4.3 alone, all were regarded as having definite or suspected organic heart disease. T-wave inversion or flattening was fairly common, having been recorded in 90 persons, especially among those aged 50 and over. The great majority of them had intermediate or minor T-wave changes (Code 5.2 and 5.3). Fifteen subjects were considered not to have organic heart disease, 22 had suspected coronary (or other) cardiac disease. Of 53 subjects with definite cardiac disease, 3 1 were classified as coronary, 16 as hypertensive disease, and the remainder as rheumatic, car pulmonale or cardiomyopathy. Left bundle-branch block encountered in 13 subjects (five men and eight women) and atria1 fibrillation (or flutter) in 14 persons (six men and eight women) were considered to indicate the presence of organic disease. Persons showing any of the previously mentioned Q or ST-T changes or exhibiting left bundle-branch block or atria1 fibrillation were classified separately (Tables 1 and 2). Age-adjusted prevalence rate of this category was 3.7% in men and 4.9% in women. C. Prevalence of coronary heart disease The age-adjusted prevalence rate of any clinical
Is3 Prevalence % Male
1 Prevalence % Female
% 1 El 16 14 12 10 8 6 4 2 0 20-29
30-39
40.49
50-59
60-69
270
Age Groups
Fig. 1. Prevalence
of coronary heart disease in Turkish by age groups.
adults
or electrocardiographic manifestation of heart disease described in Tables I and 2 were 8% and 9.6% in men and women, respectively. Since this definition probably includes a substantial number of normal subjects as well as patients with noncoronary disease such as hypertensive heart disease or cardiomyopathy, a clinically assessed prevalence of coronary heart disease was sought to be specified. These age-adjusted rates were 5.8% in Turkish men and 5% in women (P > 0.4) (Tables 1 and 2). A steep age gradient existed in the coronary heart disease prevalence in both genders, multiplying by 2.2 to five times in each life decade in women, and by roughly three times in men before declining from age 70 onwards. TABLE
3
Prevalence (in o/o) of coronary in various age groups. Age groups
All adults 20-29 years 30-39 years 40-49 years 50-59 years 60-69 years z 70 years
heart disease in Turkish
Men
Women
A
B
Rate
A
B
Rate
59 0 3 8 18 29 1
31 1 7 3 4 12 4
4 0.1 1.5 3.2 7.8 17.2 3.7
45 0 4 2 18 14 7
38 2 4 4 6 17 5
3.5 0.2 1.5 1.4 8.3 11.3 11.5
Column A, Number Column B, Number point given).
adults
of subjects with definite CHD. of subjects with suspect CHD
(one-half
28
D. Prevalence of coronary heart disease by (urban vs. rural) residence The number of persons with a definite or suspected diagnosis of coronary heart disease and their prevalences by various age groups are shown in Table 3. Age-adjusted prevalence rate of coronary heart disease in Turkish men was 5.8% in the general population, 5.7% in the urban and 6.1% in the rural population (P > 0.4). The respective rates in Turkish women were 5,6.4 and 3.5% (P c 0.06 between urban and rural sections).
E. Distribution of the presenting form of coronary heart disease Of 139 participants with clinical coronary heart disease, 87 (63%) presented angina pectoris without myocardial infarction. The latter comprised 70 subjects with typical and 34 with atypical angina (credited by one-half). In 32 patients with or without angina there was definite evidence of myocardial infarction, whereby nine persons without angina presented a history of myocardial infarction and/or Code 1. In further 10 subjects without angina Code 1 existed. Thus 27% of participants with coronary heart disease showed evidence of myocardial infarction. Silent myocardial ischemia as judged from Code 4 and/or Code 5 items in symptom-free subjects formed 7% of those having coronary heart disease, while heart failure alone was considered the presenting form of the disease in 3%.
F. Summary of distribution of some risk factors Following data are provided in order to give an idea as to the significance of underlying risk factors. Mean plasma concentrations adjusted for the age range 35-64 years were 4.8 mmoVl(185 mg/dl) in men and 5 mmol/l (192 mg/dl) in women [8]. Hypercholesterolemia (L 6/2 mmol/l (240 mg/dl) prevailed among 10% of men and 11.7% of women. Roughly two-thirds of these men and women had levels of plasma total cholesterol in excess of 6.5 mmol/l. While age-adjusted mean plasma triglyceride concentration was 1.69 mmol/l (150.2 mg/dl) in
men and 1.46 mmol/l (129.2 mg/dl) in women, hypertriglyceridemia ( 2 2.26 mmol/l, 200 mg/dl) was observed in 19.3% of men and 14.7% of women in the decades spanning the ages 40-69 years [8]. Hypertensive subjects were defined to have a systolic (~160 mmHg), diastolic (~95 mmHg) pressure, or both, or who reported to take antihypertensive medication. With an overall ageadjusted odds ratio of 2.86, hypertension constituted the most important risk factor in the sample population [ 121. It prevailed among 17% of men and 23.4% of women, when standardized for ages 35-64 years. Again when age-adjusted, 49% of Turkish men smoked daily more than 10 cigarettes, and 8% smoked up to 10 cigarettes. The respective figures for smoking women were 9% and 5.5%. Turkish women, but not men, distinctly tended to be obese: 36.5% of age-adjusted women had a body mass index ~30.0 kg/m2. Exhibiting an odds ratio of 1.76, obesity was an additional significant risk factor in women.
Discussion Coronary heart disease prevalence rates of populations obtained in epidemiologic studies have some shortcomings to be compared with each other due to a variety of reasons. Nonetheless in this study prevalences of typical and atypical angina pectoris, partial indicators of coronary heart disease, can fairly readily be compared in view of commonly adopted criteria for these symptoms. Typical angina was marginally more common in men than in women (age-adjusted 3.7% vs. 3.5%), while atypical angina was elicited twice as commonly in women (2.0% vs. 0.9% in men). In agreement with available knowledge, the relatively high prevalence of angina in women has been attributed to a lower pain tolerance [ 131, or to a greater socio-psychological acceptability in Western cultures [ 141, or to a higher proportion of women with less severe coronary heart disease [15]. The prevalence rate in Turkish women was two-thirds of that of Finnish women with respect to typical angina and one-quarter as regards atypical angina [ 151. Our rates in men were close
29
to those of Finnish men in respect with typical angina, while they constituted one-tenth as regards atypical angina [15]. Criteria for atypical angina may have been applied in a narrower sense in the present study. For further comparison one may add that typical angina prevailed in 5.5% of men aged 50-59 years in the Whitehall Study [16], in 5.8% of men aged 40-59 in the British Regional Heart Study [6] and in 5.7% of those aged 47-53 in the second Goteborg study [ 171. The prevalence of severe prolonged chest pain attack has varied widely between populations, such as for men aged 50-59 years from that of 2.4-9% [ 18,151 and data on this point might prove still less reliable as an indication of an old myocardial infarction, consequent to a lower level of health awareness among Turkish adults. Therefore, a stricter definition of a history of myocardial infarction was used here which included prolonged rest following a chest pain lasting onehalf hour or more. This provided an infarct prevalence in men that represented one-half that of typical angina. This ratio in men might somewhat underestimate the proportion of myocardial infarction in the spectrum of manifestations of coronary heart disease. Among Q/QS patterns in the ECGs, the overwhelming majority of subjects with Code 1 comprised Code 1.1 and 1.2 items. The age-adjusted prevalence of Code 1 in Turkish men was 1.8%, a figure though approaching the corresponding rate of 2.2% obtained by dstiir et al. [ 191, is half as low as that in the Framingham [20], Whitehall [ 161and Finnish [ 151 studies. The corresponding prevalence in women aged 50-59 years was approximately half that of men in three of the just mentioned studies, while Q/QS patterns prevailed in Turkish women at one-fifth the rate of male participants. The low rate of both the history of myocardial infarction and of Code 1 in Turkish women suggests that myocardial infarction predilects women much less and men more strongly than in several studies from industrialized nations. For example, among randomly selected residents of Rochester, Minnesota, among cases of coronary heart disease those with myocardial infarction constituted 54% in men and 33% in women [21]. ST depressions prevailed among the Finnish
sample population in about 3% and any T-wave changes (Code 5.1-3) in about 7.5% when similar age-adjustment is performed as in the present survey. These rates are comparable to those in several other surveys [6,8,19], while the rate for T-wave inversions among Turkish men (2.5%) was about one-third. The policy in this study of crediting half a point to asymptomatic subjects with left bundle-branch block seems supported by the experience in the Framigham study [22] in which only 11% among subjects with a newly acquired left bundle-branch block remained free of clinically apparent cardiovascular abnormalities throughout a period of 12 years’ observation: namely, since merely three of our 13 subjects with left bundle-branch block were symptom-free, 12% (1.5:13) of all those with left bundle-branch block was excluded from a diagnosis of disease. When any clinical and electrocardiographic manifestation suggestive of coronary heart disease is considered, 8% of men and 9.6% of Turkish women (age-adjusted) can be classified as such. This can be compared with the findings in the British Regional Heart Study [6] in which 24.7% of men aged 40-59 years had some evidence of coronary heart disease by survey questionnaire or ECG or both. Criteria as broad as these probably inflate the true prevalence of coronary heart disease. This is true in men, but more so in women in whom, in addition to atypical angina, minor ST-T changes are often false positively recorded [23]. Lability in the autonomic nervous system and constitutional differences in women such as a more backward directed T vector in the horizontal plane have been implicated to underly this phenomenon [24]. Therefore, a point-score system as devised herein which credits clinical and ECG findings with a low reliability by one-half point would seem to lead closer to the real prevalence of coronary heart disease. Our criteria for coronary heart disease essentially excluded isolated Codes 4.3 and 5.3 (not associated with angina) and counted half the number of persons exhibiting atypical angina or isolated electrocardiographic ‘myocardial ischemia’ of greater than minor severity. The latter approach took into account the existence of ECG changes of a noncoronary origin. Evidently, the
30
value of this assessment based on clinical judgment needs to be demonstrated in a prospective study. Nevertheless, it is provided here not as a substitute but as a supplement to the evaluation of coronary heart disease indicators obtained by standard epidemiological methods. Based on the point score, Turkish subjects, adjusted for age range 35-64 years, have a coronary heart disease prevalence of 5.8% in men and 5% in women. Overall, the age-adjusted prevalence among the urban population of 6% exceeded that of the rural population (4.8%), and this difference stemmed from women. Higher coronary heart disease prevalence data in urban residents in our study was associated with a mean total cholesterol level in the urban population (age-adjusted 193 mgdl in men vs. 182.7 mg/dl in women) which was significantly higher by 10.3 mg/dl (0.27 mmol/l) than in the rural population [8]. Though rates of symptoms and of ischemic ECG findings rose substantially with increasing age in women across all age groups and in men until age 69, lower prevalence rates of coronary heart disease were observed in men aged 70 or over implying a healthier status. Since the number of subjects comprised in this age category was comparatively low, its confidence interval is likely to be large. The low prevalence rate in the oldest men might either represent a sampling bias in this category, or it might reflect a natural selection, namely the elimination of most Turkish men with coronary disease by death before reaching the age of 70. The latter possibility is in keeping with the fact that mean age at death for Turkish men is 63.9 in contrast to 69 in women. A prevalence rate for coronary heart disease by the clinically assessed criteria in women as close to men (by merely 13% lower), or even higher than men by combined use of angina and ECG findings with standard survey methods is not an observation commonly made and needs commenting. In addition to the previously discussed factors tending to augment the coronary heart disease prevalence in women in general, a significantly higher prevalence of obesity, hypertension and of diabetes in Turkish women than in men may account for the finding of a relatively high prevalence rate of coronary heart disease in women.
In an epidemiologic cross-sectional study aimed at eliciting the prevalence of coronary heart disease angina, severe chest pain attack and ECG findings are the major tools in detection. In this survey a history of angina and/or myocardial infarction had 67% of weighting in the diagnosis, a history of bypass surgery or of angiographic tindings had a 3% weight and a similar share was by manifestations of heart failure at physical examination. The ECG provided a diagnosis in 27% of instances; of this proportion sequelae of myocardial infarction and symptomatic myocardial ischemia comprised 20%, while ‘silent’ ischemia had 7% weight. (In instances of fullillment of combined (history and ECG) half shares were allotted to both methods in the detection.) Thus two to threefold emphasis was given in this study in the diagnosis of coronary heart disease on isolated chest pain symptoms than on the presence of isolated ECG findings, an approach the correctness of which was advocated by Reunanen and associates
v51. In conclusion, this study provided a comparatively low prevalence of indicators of coronary heart disease in Turkish adults which was associated with generally low concentrations of serum total cholesterol. The prevalence rate was slightly but not significantly higher in the urban than in the rural sample population, and only marginally lower in Turkish women than in men.
Acknowledgements We appreciate the dedicated works of Dr Y. Goziikara, Dr M. Ivler, Dr Y. Karaaslan, Dr J. iSzisik, Dr F. Tabak, the coworkers in the survey teams, and that of V. Taskin and 6. oz in the data processing. The support of Professor R. &can, president of the Turkish Society of Cardiology is gratefully acknowledged. The fruitful suggestions by Professor K. Pyorala (Kuopio, Finland) in the preparation of the manuscript is deeply appreciated.
References 1
Uemura K. International trends in cardiovascular disease in the elderly. Eur Heart J 1988; 9 (Suppl D): l-8.
31 2 3
4
5
6
7
8
9
10
11
Keys A, editor. Coronary heart disease in seven countries. Circulation 1970; 41 (Suppl 1): l-195. The Pooling Project Research Group. Relationship of blood pressure, serum cholesterol, smoking habit, relative weight and ECG abnormalities to incidence of major coronary event: final report of the Pooling Project. J Chron Dis 1978; 31: 201-306. Ducimetibre P, Richard JL, Cambien F, Rakotovao R, Claude JR. Coronary heart disease in middle-aged Frenchmen: Comparisons between Paris Prospective Study, and Pooling Project. Lancet 1980; 1: 1346-50. Reid DD, Hamilton PJS, McCartney P, Rose G, Jarret RJ, Keen H. Smoking and other risk factors for coronary heart disease in British civil servants. Lancet 1976; ii: 979-83. Shaper AG, Cook DG, Walker M, MacFarlane PW. Prevalence of ischaemic heart disease in middle-aged British men. Br Heart J 1984; 51: 595-605. Bainton D, Baker IA, Sweetnam PM, Yarnell JWG, Elwood PC. Prevalence of ischaemic heart disease: the Caerphilly and Speedwell surveys. Br Heart J 1988; 59: 201-206. Onat A, Surdum-Avci G, Senocak M, Grnek E, Goziikara Y. Serum lipids and their interrelation in Turkish adults. J Epidemiol Commun Health 1992; 46: 470-476. Rose G, McCartney P, Reid DD. Selfadministration of a questionnaire on chest pain and intermittent claudication. Br J Prev Sot Med 1977; 31: 42-43. Rose GA, Blackburn H, Gillum RF, Prineas RJ. Cardiovascular Survey Methods, 2nd edn. Geneva, WHO 1982; 124-127. The WHO MONICA Project. Geographical variation in
the major risk factors of coronary heart disease in men and women aged 35-64 years. World Health Stat Q 1988; 41: 115-140. 12 Onat A, Senocak MS. Prevalence and clustering of risk factors for coronary heart disease in Turkish adults, and the associated relative risks. Arch Turk Sot Cardiol 1992; 20: 129-136.
16
17
18
19
20
21
22
23
24
Woodrow KM, Friedman GD, Siegelaub AB, Cohen MF. Pain tolerance: differences according to age, sex and race. Psychosom Med 1972; 34: 548-556. Nathanson CA. Illness and the feminine role: a theoretical review. Sot Sci Med 1975; 9: 57-62. Reunanen A, Aromaa A, Pyiirala K. Punsar S, Maatela J, Knekt P. The Social Insurance Institution’s Coronary Heart Disease Study. Helsinki: Social Insurance Institution, Finland. 1983; 120. Rose G, Baxter PJ, Reid DD, McCartney P. Prevalence and prognosis of electrocardiographic findings in middleaged men. Br Heart J 1978; 40: 636-643. Hagrnan M, Jonsson D, Wilhelmsen L. Prevalence of angina pectoris and myocardial infarction in a general population sample of Swedish men. Acta Med Stand 1977; 201: 571-577. Gyntelberg F. Physical fitness and coronary heart disease. Male residents in Copenhagen aged 40-59. Dan Med Bull 1973; 20: l-4. &tBr E, Schnor P. Jensen G. Nyboe J, Tybjaerg Hansen A. Electrocardiographic findings and their association with mortality in the Copenhagen City Heart Study. Eur Heart J 1981; 2: 317-328. Higgins ITT, Kannel WB, Dawber TR. The electrocardiogram in epidemiological studies: reproducibility, validity and international comparison. Br J Prev Sot Med 1965; 53-68. Philips SJ. Whisnant JP, O’Fallon WM, Frye RL. Prevalence of cardiovascular disease and diabetes mellitus in residents of Rochester, Minnesota. Mayo Clin Proc 1990; 65: 344-359. Schneider JF, Thomas HE Jr, Kreger BE et al. Newly acquired left bundle branch block: the Framingham study. Ann Intern Med 1979; 90: 303-310. Joy M, Trump DW. Significance of minor ST segment and T wave changes in the resting electrocardiogram of asymptomatic subjects. Br Heart J 1981; 45: 48-55. Simonson E. Differentiation between normal and abnormal in electrocardiography. St Louis. MO: C.V. Mosby Co, 1961.